Mitchell A. Schull

Hyperspectral remote sensing of foliar nitrogen content

A strong positive correlation between vegetation canopy bidirectional reflectance factor (BRF) in the near infrared (NIR) spectral region and foliar mass-based nitrogen concentration (%N) has been reported in some temperate and boreal forests. This relationship, if true, would indicate an additional role for nitrogen in the climate system via its influence on surface albedo and may offer a simple approach for monitoring foliar nitrogen using satellite data. We report, however, that the previously reported correlation is an artifact—it is a consequence of variations in canopy structure, rather than of %N. The data underlying this relationship were collected at sites with varying proportions of foliar nitrogen-poor needleleaf and nitrogen-rich broadleaf species, whose canopy structure differs considerably. When the BRF data are corrected for canopy-structure effects, the residual reflectance variations are negatively related to %N at all wavelengths in the interval 423–855 nm. This suggests that the observed positive correlation between BRF and %N conveys no information about %N. We find that to infer leaf biochemical constituents, e.g., N content, from remotely sensed data, BRF spectra in the interval 710–790 nm provide critical information for correction of structural influences. Our analysis also suggests that surface characteristics of leaves impact remote sensing of its internal constituents. This further decreases the ability to remotely sense canopy foliar nitrogen. Finally, the analysis presented here is generic to the problem of remote sensing of leaf-tissue constituents and is therefore not a specific critique of articles espousing remote sensing of foliar %N.

Monitoring water and carbon fluxes at fine spatial scales using HyspIRI-like measurements

Remotely sensed observations in the visible to the shortwave infrared (VSWIR) and thermal infrared (TIR) regions of the electromagnetic spectrum can be used synergistically to provide valuable products of land surface properties for reliable assessments of carbon and water fluxes. The high spatial, spectral and temporal resolution VSWIR and TIR observations provided by the proposed Hyperspectral - InfraRed (HyspIRI) mission will enable a new era of global agricultural monitoring, critical for addressing growing issues of food insecurity. To enable predictions at fine spatial resolution (<;100m), modeling efforts must rely on a combination of high-frequency temporal and highresolution spatial information. In this study, spatialtemporal sampling frequency is improved by employing a multi-scale and multi-sensor data fusion approach, integrating spatial detail from Landsat (30m/16 day) with the high temporal frequency of MODIS (1km/daily).

Reply to Townsend et al.: Decoupling contributions from canopy structure and leaf optics is critical for remote sensing leaf biochemistry

Townsend et al. (1) agree that we explained that the apparent relationship (2) between foliar nitrogen (%N) and near-infrared (NIR) canopy reflectance was largely attributable to structure (which is in turn caused by variation in fraction of broadleaf canopy). Our conclusion that the observed correlation with %N was spurious (i.e., lacking a causal basis) is, thus, clearly justified: we demonstrated that structure explained the great majority of observed correlation, where the structural influence was derived precisely via reconciling the observed correlation with radiative-transfer theory. What this also suggests is that such correlations, although observed, do not uniquely provide information on canopy biochemical constituents. We, therefore, disagree with the assertion in ref. 1 that we “did not provide an adequate rationale for the inference that %N and other leaf properties cannot be characterized from imaging spectroscopy”; our analysis showed precisely that. Our analysis also led to the conclusion that “NIR and/or SW broadband satellite data cannot be directly linked to leaf-level processes,” and any such link must be indirect and will be a function of structure. This is true for all wavelengths in the interval 423–855 nm (figure 7B and figure S2 in ref. 3), not primarily for the 800- to 850-nm spectral band, as misstated in ref. 1. None of the leaf biochemical constituents can be accurately estimated without accounting for canopy structural effects.

A global map of rainfed cropland areas at the end of last millennium using remote sensing and geospatial techniques

Rainfed agriculture plays a critical role in most part of the tropics and subtropics of the world. Eighty percent of the agricultural land worldwide is under rainfed agriculture; and significant proportion of rural economy still depends on rainfed agriculture with characteristically low yield levels. In this context the International Water Management Institute (IWMI) produced the first satellite sensor based Global map of rainfed cropland areas at 10Km resolution (GMRCA10Km). The study used a mega-file of 159 global data layers involving the AVHRR and SPOT time-series, GTOPO30 DEM, mean monthly rainfall, and forest cover. A suite of innovative techniques were developed that begins with the image segmentation, quantitative spectral matching techniques (SMTs) and spectral correlation similarity (SCS R2). The SCS was found to be the most useful technique in grouping identical classes. Mixed classes were resolved using a decision trees, time series plots, and principal component analysis algorithms. A wide array of groundtruth data, and high-resolution images were used to identify and label classes. The outcome was the GMRCA10Km estimated to be 1.75 billion hectares for the main cropping period. The sub-pixel areas (SPAs) of GMRCA10Km provide more realistic estimates of the actual area cultivated unlike the full pixel areas (FPAs) often calculated from the raster datasets. Three distinct GMRCA10Km maps have been produced: viz., Aggregated 7-class, Dis-aggregated 18-class and Generic 255-class. The aggregated classes will suffice for broad range of users at global level. The GMRCA10Km product line consists of maps, images, area calculations, snap-shots, class characteristics, and animations.

Reply to Ollinger et al.: Remote sensing of leaf nitrogen and emergent ecosystem properties.

Various physical, chemical, and physiological processes, including canopy structure, impact surface reflectance. Remote sensing aims to derive ecosystem properties and their functional relationships, given these impacts. Ollinger et al. (1) do not distinguish between the forward and inverse problems in radiative transfer and, hence, misrepresent our results (2). The authors also suggest our conclusions are based on a subset of data from ref. 3, which is not the case.

The application of spectral invariants for discrimination of crops using CHRIS-PROBA data

Numerous studies have demonstrated the ability of hyper-spectral data to discriminate crop types, however most methods rely on empirical data and are therefore site specific. In this brief proceeding we provide a physically based approach for separation of crop types using multiangle hyperspectral data. We use the radiative transfer theory of spectral invariants which allows for the parameterization of the canopy reflectance into two spectrally invariant and structurally varying parameters — recollision and escape probabilities. The spectral invariant parameters are retrieved from the CHRIS/PROBA multiangle hyperspectral sensor. We present the spectral invariant parameters in spectral invariant space. The horizontal axis provides information about macro scale features such as plant shape and size as well as ground cover. The vertical axis provides information about microscale features such as leaf density as well as portion of sunlit to shaded leaves. These features allow for the natural separation of crops. In addition we illustrate the potential for further separation of crop types based on angular information. Results suggest that multiangle information is important for canopies with similar structural features in the nadir direction.

Canopy spectral invariants for remote sensing of canopy structure

The concept of canopy spectral invariants expresses the observation that simple algebraic combinations of leaf and canopy spectral reflectances become wavelength independent and determine two canopy structure specific variables - the recollision and escape probabilities. The recollision probability (probability that a photon scattered from a phytoelement will interact within the canopy again) is a measure of the multi-level hierarchical structure in a vegetated pixel and can be obtained from hyperspectral data. The escape probability (probability that a scattered photon will escape the vegetation in a given direction) is sensitive to canopy geometrical properties and can be derived from multi-angle spectral data. The escape and recollision probabilities have the potential to separate forest types based on crown shape and the number of hierarchical levels within the landscape. This paper introduces the concept and demonstrates how this approach can be used to monitor forest structural parameters with multi-angle and hyperspectral data.

On the relationship between nominal light use efficiency and leaf chlorophyll

A light-use-efficiency (LUE)-based model of canopy resistance has been embedded into a thermal-based Two-Source Energy Balance (TSEB) model to facilitate coupled simulations of transpiration and carbon assimilation. The model assumes that deviations of the observed canopy LUE from a nominal stand-level value (LUEn - typically indexed by vegetation class) are due to varying conditions of light, humidity, CO2 concentration and leaf temperature. The deviations are accommodated by adjusting an effective LUE that responds to the varying conditions. We investigate the feasibility of leaf chlorophyll (Cab) to capture these variations in LUEn using remotely sensed data. To retrieve Cab from remotely sensed data we use REGFLEC, a physically based tool that translates at-sensor radiances in the green, red and NIR spectral regions from multiple satellite sensors into realistic maps of LAI and Cab. Initial results show that using Cab to estimate LUE allows for improved flux estimates over a soybean field in Iowa. The improved results indicate the necessity of a varying LUE during times of stresses induced by the environment. The results also indicate that using remotely sensed Cab to estimate LUE will allow for more accurate estimates of fluxes over space and time.

An integrated approach for high spatial resolution mapping of water and carbon fluxes using multi-sensor satellite data

In the last years, modeling of surface processes - such as water, energy and carbon budgets, as well as vegetation growth- seems to be focused on integrated approaches that combine aspects of hydrology, biology and meteorology into unified analyses. In this context, remotely sensed data often have a core role due to the cross-cutting impact of this novel source of spatially distributed information on all these research areas. However, several applications - such as drought monitoring, yield forecasting and crop management - require spatially detailed products at sub-field scales, which can be obtained only with support of adequately fine resolution remote sensing data (< 100 m). In particular, observations in the visible to the near infrared (VIS/NIR) spectral region can be used to derive biophysical and biochemical properties of the vegetation (i.e., leaf area index and leaf chlorophyll). Complementarily, the thermal infrared (TIR) signal provides valuable information about land surface temperature, which in turn represents an accurate proxy indicator of the subsurface moisture status by means of surface energy budget analysis. Additionally, the strong link between crop water stress and stomatal closure allows inference of crop carbon assimilation using the same tools. In this work, an integrated approach is proposed to model both carbon and water budgets at field scale by means of a joint use of a thermal-based Two Source Energy Budget (TSEB) model and an analytical, Light-Use-Efficiency (LUE) based model of canopy resistance. This suite of models allows integration of information retrieved by both fine and coarse resolution satellites by means of a data fusion procedure. A set of Landsat and MODIS images are used to investigate the suitability of this approach, and the modeled fluxes are compared with observations made by several flux towers in terms of both water and carbon fluxes.

Modeling recollision and escape probabilities using the stochastic radiative transfer equation

The spectral invariants of vegetation canopy convey a lot of information on the canopy structure at hierarchical levels. Recent findings of the wavelength independent and scale independent variable — the ratio between the directional escape and the total escape probability — show that it does dependent on the selection of reference leaf albedo in getting correct reflectance values and can be treated as the identifier of macro scale canopy structure (foliage density, aspect ratio, ground cover, tree shape). In order to better utilize this variable in the retrieval algorithm for 3D canopy structure. Model simulation based on the stochastic radiative transfer equation is used to test the sensitivity of this variable to the structural parameters. Hyperspectral and multi-angle data are simulated and analyzed.